Tensor accelerators now represent a growing share of compute resources in modern CPUs and GPUs. However, they are hard to program, leading developers to use vendor-provided kernel libraries that support tensor accelerators. As a result, the usage of tensor accelerators is limited to the provided interface, mainly designed for traditional ML and scientific computing workloads.In this paper, we show that tensor accelerators can improve the performance of applications beyond simple variants of MatMul. For example, many image processing pipelines are linear transformations over matrices in disguise and can therefore utilize such specialized hardware. This is nonetheless hindered by the difficulties in programming tensor accelerators. We tackle this problem with compiler-based techniques. We use the Halide user-schedulable language and express operations as Halide algorithms succinctly. To this end, we implement a flexible tensor instruction selector based on equality saturation. The tensor instruction selector supports both CPU- and GPU-attached tensor accelerators and works with existing scheduling operations (e.g., producer-consumer fusion). Together, this enables developers to write diverse accelerator-leveraging applications in a few dozen lines.Using our system, we demonstrate the potential of tensor accelerators beyond their traditional domains. We implement several image processing pipelines (e.g., filtering, resampling, and denoising) in our system and evaluate them against non-accelerator-leveraging baselines. We show that these pipelines can achieve significant speedups. For example, a downsampling routine is sped up by 6.1× by utilizing Tensor Cores on an Nvidia RTX 4070 GPU.

Abstract Parametric CAD systems use domain‐specific languages (DSLs) to represent geometry as programs, enabling both flexible modeling and structured editing. With the rise of large language models (LLMs), there is growing interest in generating such programs from natural language. This raises a key question: what kind of DSL best supports both CAD generation and editing, whether performed by a human or an AI? In this work, we introduce AIDL, a hierarchical, solver‐aided DSL designed to align with the strengths of LLMs while remaining interpretable and editable by humans. AIDL enables high‐level reasoning by breaking problems into abstract components and structural relationships, while offloading low‐level geometric reasoning to a constraint solver. We evaluate AIDL in a 2D text‐to‐CAD setting using a zero‐shot prompt‐based interface and compare it to OpenSCAD, a widely used CAD DSL that appears in LLM training data. AIDL produces results that are visually competitive and significantly easier to edit. Our findings suggest that language design is a powerful complement to model training and prompt engineering for building collaborative AI–human tools in CAD. Code is available at https://github.com/deGravity/aidl .

Modern vector processors support a wide variety of instructions for fixed-point digital signal processing. These instructions support a proliferation of rounding, saturating, and type conversion modes, and are often fused combinations of more primitive operations. While these are common idioms in fixed-point signal processing, it is difficult to use these operations in portable code. It is challenging for programmers to write down portable integer arithmetic in a C-like language that corresponds exactly to one of these instructions, and even more challenging for compilers to recognize when these instructions can be used. Our system, Pitchfork, defines a portable fixed-point intermediate representation, FPIR, that captures common idioms in fixed-point code. FPIR can be used directly by programmers experienced with fixed-point, or Pitchfork can automatically lift from integer operations into FPIR using a term-rewriting system (TRS) composed of verified manual and automatically-synthesized rules. Pitchfork then lowers from FPIR into target-specific fixed-point instructions using a set of target-specific TRSs. We show that this approach improves runtime performance of portably-written fixed-point signal processing code in Halide, across a range of benchmarks, by geomean 1.31x on x86 with AVX2, 1.82x on ARM Neon, and 2.44x on Hexagon HVX compared to a standard LLVM-based compiler flow, while maintaining or improving existing compile times.

Parametric CAD models encode entire families of shapes that should, in principle, be easy for designers to explore. However, in practice, parametric CAD models can be difficult to manipulate due to implicit semantic constraints among parameter values. Finding and enforcing these semantic constraints solely from geometry or programmatic shape representations is not possible because these constraints ultimately reflect design intent. They are informed by the designer’s experience and semantics in the real world. To address this challenge, we introduce ReparamCAD, a zero-shot pipeline that leverages pre-trained large language and image model to infer meaningful space of variations for a shape We then re-parameterize a new constrained parametric CAD program that captures these variations, enabling effortless exploration of the design space along meaningful design axes. We evaluated our approach through five examples and a user study. The result showed that the inferred spaces are meaningful and comparable to those defined by experts. Code and data are at: https://github.com/milmillin/ReparamCAD.

After decades of progress, database management systems (DBMSs) are now the backbones of many data applications that we interact with on a daily basis. Yet, with the emergence of new data types and hardware, building and optimizing new data systems remain as difficult as the heyday of relational databases. In this paper, we summarize our work towards automating the building and optimization of data systems. Drawing from our own experience, we further argue that any automation technique must address three aspects: user specification, code generation, and result validation. We conclude by discussing a case study using videos data processing, along with opportunities for future research towards designing data systems that are automatically generated.

Instruction selection, whereby input code represented in an intermediate representation is translated into executable instructions from the target platform, is often the most target-dependent component in optimizing compilers. Current approaches include pattern matching, which is brittle and tedious to design, or search-based methods, which are limited by scalability of the search algorithm. In this paper, we propose a new algorithm that first abstracts the target platform instructions into high-level uber-instructions, with each uber-instruction unifying multiple concrete instructions from the target platform. Program synthesis is used to lift input code sequences into semantically equivalent sequences of uber-instructions and then to lower from uber-instructions to machine code. Using 21 real-world benchmarks, we show that our synthesis-based instruction selection algorithm can generate instruction sequences for a hardware target, with the synthesized code performing up to 2.1x faster as compared to code generated by a professionally-developed optimizing compiler for the same platform.

This paper presents Dexter, a new tool that automatically translates image processing functions from a low-level general-purpose language to a high-level domain-specific language (DSL), allowing them to leverage cross-platform optimizations enabled by DSLs. Rather than building a classical syntax-driven compiler to do this translation, Dexter leverages recent advances in program synthesis and program verification, along with a new domain-specific synthesis algorithm, to translate C++ image processing code to the Halide DSL, while guaranteeing semantic equivalence. This new synthesis algorithm scales and generalizes to much larger and more complex functions than prior work, including the ability to handle tiling, conditionals, and multi-stage pipelines in the original low-level code. To demonstrate the effectiveness of our approach, we evaluate Dexter using real-world image processing functions from Adobe Photoshop, a widely used multi-platform image processing program. Our results show that Dexter can translate 264 out of 353 functions in our test set, with the original implementations ranging from 20 to 150 lines of code. By leveraging Halide's advanced auto-scheduling capabilities, we get median speedups of 7.03× and 4.52× for Dexter-translated functions as compared to the original implementations on Intel and ARM architectures, respectively.

MapReduce is a popular programming paradigm for developing large-scale, data-intensive computation. Many frameworks that implement this paradigm have recently been developed. To leverage these frameworks, however, developers must become familiar with their APIs and rewrite existing code. We present Casper, a new tool that automatically translates sequential Java programs into the MapReduce paradigm. Casper identifies potential code fragments to rewrite and translates them in two steps: (1) Casper uses program synthesis to search for a program summary (i.e., a functional specification) of each code fragment. The summary is expressed using a high-level intermediate language resembling the MapReduce paradigm and verified to be semantically equivalent to the original using a theorem prover. (2) Casper generates executable code from the summary, using either the Hadoop, Spark, or Flink API. We evaluated Casper by automatically converting real-world, sequential Java benchmarks to MapReduce. The resulting benchmarks perform up to 48.2x faster compared to the original.

Parallelizing of software improves its effectiveness and productivity. To guarantee correctness, the parallel and serial versions of the same code must be formally verified to be equivalent. We present a novel approach, called GRASSP, that automatically synthesizes parallel single-pass array-processing programs by treating the given serial versions as specifications. Given arbitrary segmentation of the input array, GRASSP synthesizes a code to determine a new segmentation of the array that allows computing partial results for each segment and merging them. In contrast to other parallelizers, GRASSP gradually considers several parallelization scenarios and certifies the results using constrained Horn solving. For several classes of programs, we show that such parallelization can be performed efficiently. The C++ translations of the GRASSP solutions sped performance by up to 5X relative to serial code on an 8-thread machine and Hadoop translations by up to 10X on a 10-node Amazon EMR cluster.

In this demonstration we will showcase CASPER, a novel tool that enables sequential data-intensive programs to automatically leverage the optimizations provided by parallel data processing frameworks. The goal of CASPER is to reduce the inertia against adaptation of new data processing frameworks---particularly for non-expert users---by automatically re-writing sequential programs written in general purpose languages to the high-level DSLs or APIs of these frameworks. Through CASPER's browser-based interface, users can enter the source code of their Java applications and have it automatically retargeted to execute on Apache Spark. In our interactive presentation, we will use CASPER to optimize sequential implementations of data visualization programs as well as image processing kernels. The optimized Spark implementations along with the original sequential implementations will then be executed simultaneously on the cloud to allow the demo to audience compare the runtime performances and outputs in real-time.